292
ANALYTICAL CHEMISTRY
dlthough accuracy rather than speed has been of primary concern in this investigation, automatic titrations with the present instrument absorb less operator time than the same titrations by the manual method. I t is not feasible to increase the delivery rate much above 4 cc. per minute in most titrations, because overrunning is inevitable if the equivalence point is approached too rapidly. If greater speed is desired it should be obtained by adding most of the titrant at a rdpid rate and then finishing the titration more slowly as in manual methods. This could be achieved relatively simply by employing a syringe(hire motor whose speed can be varied over a three- or fivefold range by a series resistance. -\permanent resistor could be placed in series with the motor t o establiih the slow speed, and a second mercury switch operated like sn-itch S,by the recording potentiometer n-ould be connected acro.ss this resistor in the normally closed position to short out the resistor for high-speed delivery. This second srvitch would be set to open slightly before the equivalence point, so that most of the titrant would be delivered rapidly and the equivalence pbint n-ould finally be approached at a slower speed just as in manual titrations. The use of an expensive recording potentiometer is, of course, not necessary, since the titration curve does not have to be recorded once it5 characteristics are k n o ~ n . .In? commercial potentiometer controller should be satLfactory, provided it has rapid response, proper range, and input characteristies (high impedance) suitable for electrode syqtenw. -I relatively inespensive electronic trigyer circuit ha.: been developed in this
laboratory to replace the potentiometer recorder, and this instrument will be described in a following paper. LITERATURE CITED
(1) Berhenke, L. F., and Britton, E. C., Ind. Eng. Chem., 38, 544 (1946). (2) Dubois, P., Compt. rend., 194,1162 (1932). (3) Furman: N. H., J . Am. Chem. SOC.,50, 1675 (1928). (4) Gutaeit, G., and Devaud, C:, Arch. sei. phys. nat. (5), 11, 67 (1929); Chem. Z e n t r . , 1929,1715. (5) Kolthoff, I. M., and Furman, N. H., “Potentiometric Titiations,” New York, John Wiley & Sons, 1931. (6) Kolthoff, I. M.,and Lingane, J. J., J . Am. Chem. Soc., 58, 1524, 1528, 2457 (1936). (7) Kordataki, W., and Wulff, P., Z . anal. Chem., 89, 241 (1932):, Chem. Fabrik, 3, 329, 342 (1930). (5) Lingane, J. J., IND. ESG. CHEM.,ASAL. ED., 18, 734 (1946). (9) MacInnes, D. A , , and Belcher, D., J . Am. Chem. SOC.,55, 2630 (1933). (10) Mtlller, E., “Elektrometrische Massanaiyse,” Dresden und Leipaig, Theo. Steinkopff, 1942. (11) Mtiller, E., 2. angew. Chem., 41, 1176 (1928). (12) Sottingham, W.B., J . Franklin Inst., 209, 287 (1930). (13) Shedlovsky, T., and MacInnes, D. A , , J . Am. Chem. Soc., 57, 1705 (1935). (14) Shenk, JV. E., and Fenwick, F., ISD. ESG. CHEM.,ANAL.ED., 7, 194 (1935). (15) Yogels, H., Bull. sci. acad. roy. Belg. ( 5 ) , 19,452 (1933). (16) Khitnah, C. H., IND. ENG.CHEM.,ANAL.ED., 5 , 352 (1933). (17) Willard, H. H., andYoung. P., Ind. Eng. Chem., 20,972 (1928). RECEIVED J u l y 12, 1947.
Improvements in Rubber Testing in the Government Synthetic Rubber Program ROBERT D. STIEHLER, National Bureau of Standards, Vashington, D. C. RICHARD W. HkCKETT, Ofice of Rubber Reserve, Reconstruction Finance Corporation, Washington, D. C. In the government synthetic rubber program, the precision and accuracy of testing rubber have been increased severalfold. The vaIues of the standard deviations obtained at the beginning of the program in 1943 and 1944 were two to fifteen times as large as those obtained at present. The improvement in testing resulted from standardization and improvement of technique, refinement of existing methods
T
HE purpose of this paper is to report the improvements in precision and accuracy of rubber testing achieved in the government synthetic rubber program during the past four years. These improvements have resulted largely from the activities of the Office of Rubber Reserve and its associated committees and synthetic rubber plants. I n 1943, the Office of Rubber Director, \Tar Production Board, and the Rubber Reserve Company (now Office of Rubber Reqerve, Reconstruction Finance Corporation) decided that to meet the rubber crisis the government plants should produce a single, general-purpose synthetic rubber of the butadienestyrene type, which was designated GR-9. This decision necessitated that the production from all of the fifteen government GR-S plants be interchangeable and uniform in quality. To attain this goal, the Office of the Rubber Director and Rubber Reserve ( 2 ) jointly established a Committee on Specifications for Synthetic Rubbers, which held its first meeting on May 12, 1943. Besides representatives of these two agencies, representatives of the rubber compounders, the operators of the synthetic rubber plants, and the Sational Bureau of Standards were members of this committee.
of test, and development of improved methods of test. Each of these approaches is reviewed. The results which have been realized in this program indicate that a high degree of precision can be obtained if care is taken to control test conditions and proper technique is employed. Such precision has been attained without a significant increase in man. hours or cost of equipment. The Committee on Specifications for Synthetic Rubbers recognized a t its first meeting the poor reproducibility of testing then existing within and among the rubber laboratories. It also realized that a uniform product could not be achieved unless the precision and accuracy of testing were improved. Consequently, a Subcommittee on Physical Testing Methods was established, which in May 1944 became the Subcommittee on Test Methods ( 8 ) and thereafter embraced both chemical and physical testing. I n order to reduce the variability in rubber testing, the activities of these committees and the Office of Rubber Reserve Ivere directed toward: 1. Standardization and improvement of technique 2. Refinement of existing methods of test 3. Development of improved methods of test STANDARDIZATION AND IMPROVEMENT OF TECHNIQUE
When the synthetic rubber plants began operation, only a few of the laboratory supervisory personnel had had any experience in the testing of rubber. Therefore, before the plants began production, the laboratory personnel were given several
V O L U M E 20; NO. 4, A P R I L 1 9 4 8 Table I.
Plant
Acetone Extract
%
Analysis of Blended GR-S Sample H by li Synthetic Rubber Laboratories
Fatty Acid
%
Viscosity Compound hlL. b XIL. b
GR-S
Soap
%
7.7 4.8 0.15 6.9 4.9 0.15 4.5 0.20 6.7 3 7.5 5.3 0.15 4 4.7 0.10 7.3 5 4.8 0.10 7.1 6 4.6 0.15 7.2 7 7.0 5.0 0.10 8 7.0 4.6 0.10 9 6.0 4.2 0.15 10 4.8 0.10 6.6 11 5.2 0.10 6.8 12 7.0 5.0 0.15 13 4.7 0.10 6.6 14 5.0 0.15 6.8 15 4.9 0.10 7.0 16 7.0 4.8 0.15 High 5.3 0.20 7.7 Low 4.5 0.10 6.0 7.0 4.8 0.15 Average Standarddeviation 0.38 0.25 0.036 a 50-minute cure. 411 cures a t 292’ F. b ML., Mooney viscometer with large rotor. 0 1 2
293
Bsh
% 1.00 1.00 1.15 1.15 1.00 1.00 1.00 0.90 1.05 0.60 1.05 1.00 0.90 1.05 1.00 1.05 1.00 1.15 0.60 1.00 0.117
46.5 47.0 50.0 51.0 49.0 52.0 47.0 48.0 50.0 48.0 50.0 52.0 48.0 49.0 49.0 55.0
57.0 49.0 65.0 62.0 60.0 55.0 58.0 65.0 61.0 58.0 66.0 55.0 57.0 62 .O 60.0 66.0
55.0 46,5 49.5 2.16
66.0 49.0 59,s 4.52
...
...
neeks’ training to familiarize them with the various techniques. Sonetheless, it was a common experience to have extremely erratic results reported from a single laboratory. Liken-ise, marked variation was noted among the results of tests from the different laboratories on a single sample of blended rubber. The results obtained on one such sample are shown in Table I. These results, obtained after nearly one year of*operation, made it apparent that testing variability had to be reduced if a uniform product interchangeable from plant to plant were ever to be realized. One of the most important steps taken to standardize testing was the establishment of reference lots of rubbers for daily tests by each analyst. The primary requisite for a reference lot is uniformity] both from bale to bale and within each bale. T o prepare such a lot of GR-S, a designated plant segregates during normal operation sufficient latex for 50,000 to 60,000 pounds of dry rubber. This latex is thoroughly agitated and then coagulated and dried on a single finishing line. Three hundred bales (20,000 to 25,000 pounds) from the central portion of the run are selected for the reference lot. Such a lot is uniform with respect to physical properties, but varies slightly in ash content and other chemical properties because of slight variations in the washing process. Consequently, for the chemical analyses, three of the bales are thoroughly blended on a large mill. This blended material becomes the reference lot in control testing for the chemical properties. Each plant receives a sufficient quantity of the blended material and enough bales of the unblended portion of the lot for a daily appraisal of testing for each property over a period of six months. From each bale used by the plants, a 2-pound sample is taken a t the time of production and sent to the National Bureau of Standards, together with a sample of the blended material. Values for the several chemical and physical properties are established from the results obtained by the National Bureau of Standards and the plants during the first month’s testing of the reference lot. Reference lots of GR-I and GR-11 are smaller in size, but the same care is taken to ensure uniformity. By a comparison of the results of the daily tests made by each analyst with the established values the plant laboratories can determine whether improper techniques are being employed by a particular analyst or systematic errors are being introduced through faulty equipment. Although the reference lot serves admirably t o indicate the quality of testing in a particular plant it is doubtful whether testing errors would have been eliminated
Tensilea Strength L b . / s q . anch. 3040 3000 3045 3025 2920 3250 2940 3020 3010 2890 2790 3070 3110 3060 2980 2970 3075 3250 2890 3010 Q7
Ultimate Elongations
53 610 625 610 600 600 580 625 640 605 610 580 625 575 610 615 590 600 640 575 605 17
Modulus at 300% Elongation 25-minute 50-minute 90-minute cure cure cure Pounds per square inch 480 1080 1460 1525 490 1030 560 1015 1385 1600 550 1125 450 990 1380 700 1210 1500 510 925 1365 550 1010 1290 1510 600 Ill5 1400 465 1005 500 1065 1490 1310 455 985 1440 500 1115 540 101.5 1360 510 I020 1320 575 1040 1380 590 1100 1410 700 1210 1600 450 925 1290 1420 530 1050 82 61 66
without a simultaneou; educational and coordination piogram ( 3 ) . This program i- under the supervision of the producttesting and quality-control group of the Office of Rubber Reserve. Members of this group, thoroughly familiar lvith the equipment and techniques involved in the several methods of test, frequently visit the plant laboratories to instruct and advise laboratory personnel. I n addition, a field representative is assigned to each plant. These representatives assist in standardizing the testing techniques as well as verifying the quality of production. As an illustration of such standardization, the experience n-ith the determination of ash may be cited. Although the method had been well established, the tendency of the plant laboratories in the early part of the program was to obtain low and erratic results for ash content. An examination of the techniques employed revealed that the temperatures of the muffle furnaces were variable. When the temperature exceeded the specification limits, sodium chloride and other salts were vaporized or decomposed. I n order to obtain reproducible and accurate results, the temperature limits for ashing were changed from 550 * 50 ’ C. to 550’ * 25 O C. and temperature controllers were installed on the muffle furnaces. As a further precaution, the use of aluminum crucibles was recommended. These crucibles distort and melt if the temperature is appreciably above that permitted by the specification. They also have other advantages over porcelain crucibles, such as lower cost, lighter and less variable weight, and more rapid heating and cooling. These improvements in technique have eliminated over 80 per cent of the testing error. O
REFINEMENT OF EXISTING ANALYTICAL METHODS
In standardizing techniques in the government synthetic rubber plants, many instances were noted where the procedures were not sufficiently specific to yield reproducible results. When this occurred, the Su?xoniniittee on Test Xethods assigned to one or more laboratoiies the task of investigating the factors n-hich Fere causing the erratic results. These studies led to a refinement of the methods in the specifications. Since the refinements made in the methods of test have been incorporated in the “Specifications for Government Synthetic Rubbers” (4), it will suffice t o confine the discussion here t o one typical caseextraction of GR-S. In the early specifications, acetone was designated as the solvent to be used in the determination of extract and fatty acid in the GR-S. It was soon noted, however, that consistently high and erratic results for extract were obtained using acetone. On investigation it was found that the high value for extract resulted
.
A N A L Y T I C A L' C H E M I S T R Y
294
mination of soap content, on the same sample used t o determine fattyiteid content. Some of the other refinemcnt,s and improvoment,s in test,ing procedures made during the.past four years (4,5, 6, 7) me the following:
from the polymcriztttion of the acetone, owing t,o the presencc
of soap in the GR-S. The National Bureau of Standards investigated several solvents as a replacement of acetone, and finally recommended t.he use of s n azeotrope of ethanol and toluene. This azeotrope s ~ e l l the s rubber, thereby permitting the solvent to penetrate and extract the sample more readily. This solvent d m extracts the soap, and t,hus permlts the drtei -
fatty aoid saap titration by m-cresol purple as end-point indicator.
Tablc 11. I m p r o v e m e n t in Precision of Testing in a Typical Plant No. of tests 30 30 30 30
(Summary o i daily tests made on reference ramplm) Allgust 1943 Standard Xo. of
High
LOW
1.61 9.08 4.81 0.64
0.36 5.71 3.33 0.36
.....
.....
.....
.....
visoosity o i compound
viaeoaity of
GR-S
ML. ,o .............
0.24 0.72 0.30 0.07
....
.....
....
.....
57 73
53 60
55 66
30 30
3180 695
2525 61s
2840
Tenrile sfrength
deviation
0.94 7.58 4.06 0.47
30 30
30 380 530 800 1030 30 1045 137i 30 Acetone used 8s solvent in 1943, ethanol-toluene sweotroue in 1947.
ML.
Mehn
65;
1.06 2.86 129 20
446 40 915 65 1230 71 b 50-minute oure.
Ultimate e1ongstiou
'
June 1947
Standard deviation
LOW
Mean
0.50 1.42 4.73 0.26 1.57 23.1
0.34 7.21 4.60 0.21 1.42 22.9
0.40 7.26 4.645 0.230 1.49 22.98
0.04 0.06 0.03 0.01 0.04 0.06
51.0 59.0
50.0 57.0
50.65 58.10
0.34 0.52
tests 23 23
High
23 23 23 23
23 23 33 33
3540 690
3020 610
3233 656
118 19
33 33 33
515 1110 1510
440 1030 1420
469 1076 1477
20
Allcuresat 292" F.
Stress a t 300 per oent elongation 50-minute cure 90-minute cure
25-lz>in,.te CUTe
._
5
0
.5:..:
..
x
-
55
GR.S
19 22
I944
_,
Figure 1. Improvement in Precision and Accuracy of Testing in GR-S Plant Laboratories frbm 1944 to 1947
295
V O L U M E 20, N O . 4, A P R I L 1 9 4 8 2. Modified ash procedure to permit rapid ashing without loss of precision and accuracy. 3. Improved design of hlooney viscometer and refined procedure for determining Mooney viscosity. 4. Established standard compounding ingredients. 5. Established detailed procedure for mixing and aging compound before curing. 6 . Improved procedure for standardizing curing temperatures. 7. Improved specification for specimen die and procedure for preparation of test specimens. 8. Improved procedure for testing specimens and calibration of testing machines. DEVELOPMEKT OF IMPROVED ANALYTICAL METHODS I n many instances test methods for properties which are unique for synthetic rubber-e.g., bound styrene in GR-S-were rrquired, and existing procedures sometimes were not satisfactory. In these cases, the Subcommittee on Test hIethods undertook the development of new ones. Included among those developed are the following: 1. Mill method for volatile matter content. 2. Pyrolysis method for determination of carbon black in GRS blacks. 3. Ultraviolet absorption method for determination of stabilizers. 4. Refractive-index method for determination of bound rtyrene. 5. Strain test for determination of stress-strain propertiea. Since the first three methods are given in the “Specifications for Government Synthetic Rubbers” ( d ) , only the last two methods will be discussed here. Refractive Index Method for Determination of Bound Styrene. At the beginning of the government synthetic rubber program there was no suitable procedure available for the control testing of bound styrene in GR-S. To the National Bureau of Standards was assigned the task of developing a rapid and accurate test, and as a result the refractive-index method ( 7 ) was established. A relationship between bound styrene and refractive index was determined by carbon-hydrogen analyses and a conveision table was prepared. Briefly, the method consists of extracting nonhydrocarbon constituents from a small sample o f GR-S, pressing the extracted sample into a thin sheet, placing this sheet on the prism of an Abbe-type refractometer, and measuring the index of refraction. By use of tables, the refractive index is converted to per cent of bound styrene. Strain Test for Determination of Stress-Strain Properties. It was found early in the program that a significant part of the variability in testing was introduced by the testing equipment, Since the latter part of 1944, the Sational Bureau of Standards has been developing a strain test ( 5 )to overcome this difficulty. This test is based on the measurement of strain a t a fixed stress instead of the measurement of stress a t a specified elongation during the extension of the specimen. A semiautomatic tester ( 1 ) for this purpose has been constructed which reduces t hr man-hour requirements below those required in the usual tests. The variations introduced by this method of measurement are negligible in comparison with those introduced in the process of mixing and curing. An additional advantage of this test results from the observation t h a t strain a t a fixed stress decreases with time of cure, according to the lans of a secondorder reaction. Thus, it is possible to calculate from strain test d a t a three rulcanization parameters-scorch time, rate of cure, and structure rigidity. DISCUSSION The end result of these modes of attack can be seen in Table I1 and Figure 1. Table I1 shows the reduction in the variability of testing during the past four years within a single plant for each of the chemical and physical properties. Figure 1, which
Table 111. Precision of Testing at GR-S Plant Laboratory Operated by National Synthetic Rubber Corporation (February 1947) Property
Ash, % E-T-A extract, % F a t t y acid, % Soap, % Stabilizer, % Viscosity Bound styrene, of GR-S %
Act u a1 Standard Deviation Range of of Test Valuesa Test Values0 0.06 0.02 0.15 0.04 0.05 0.02 0.04 0.01 0.08 0.02 0.12 0.06 1.5 0.37 2.5 0.61 105 37 25 7
Viscosity of compound Tensile strength, lb./sq. inch Cltimate elongation, lb./sq. inch Stress a t 300% elongation 25-minute cure, lb./sq. inch 50-minute cure, lb./sq. inch 90-minute cure, lb./sq. inch Values obtained from 19 daily tests for each Q
11 12 16
:; 45
property.
summarizes over 10,000 test values obtained in the laboratories of 16 GR-S plants during control testing of reference lot X-55 GR-S in 1944 and X-346 GR-S in 1947, shows for each of the physical properties the range of the individual values and the mean values with respect to the over-all average. The over-all length of the bars represents the range of the individual values or precision of testing. The white segment in the bar represents the deviation of the mean value found by the plant from the average of all values-that is, the accuracy of testing. Although there has been an improvement in precision of testing for all properties, the improvement is much more striking for some than for others. There has been an improvement in accuracy of testing for viscosity both before and after compounding and for stress a t 3007, elongation, but little or no improvement for tensile strength and ultimate elongation. Konetheless the 1947 results represent a precision and accuracy of testing among plant laboratories which is probably the best performance ever experienced in the rubber industry. The performance of the plant laboratory operated by the National Synthetic Rubber Corporation, Louisville, Ky., prior t o the time it was placed in standby condition, surpassed even the best single-plant performance shown in Figure 1. The standard deviations and actual ranges of values obtained in daily tests of X-346 GR-S during February 1947 are shown in Table 111. These results indicate that even the present precision of testing in the synthetic rubber plants can be improved. I n physical testing, a large part of the error still remaining is caused by variations in atmospheric conditions of temperature and humidity during mixing and aging of the unvulcanized stock and variations in the testing machine. By use of the strain tester (1) recently developed and with air-conditioned laboratories, it should be possible to test rubber with a precision approaching that of testing simple chemical compounds. The pretext that rubber is variable in its properties can no longer be used t o account for erratic test data. The uniformity of current synthetic rubber production was made possible by improvements in the precision and accuracy of testing rubber. These improvements have been realized with little increase in cost of equipment or labor, since the man-hours required to follow prescribed procedures properly and carefully are no more than those required to test in a haphazard and careless manner. The little extra labor required t o maintain testing equipment in good operating condition and to test daily a reference control sample for standardization purposes is well repaid in reliability of results. Most rubber manufacturers today accept the test data from the synthetic rubber plants without question. CONCLUSIONS The precision of testing in the synthetic rubber plants indicates that a rubber laboratory can readily attain results on re-
ANALYTICAL CHEMISTRY
296
vidual who made contributions, the authors wish to mention the following organizations, of which these individuals are members:
Table IF-. Precision of Testing Standard Deviation 0.02 0.12 0.05 0.01 0.02
Proi~erty Ash, R Extract (E-T-A), % F a t t y acid, % Soap, % Stabilizer, yc Viscosity GR-S, N L , , 212O F. a t 4 niin. Compound, N L . , 212O F.a t 4 Inin. Properties of rulcanizate Tensile strength, lb./sq. inch Ultimate elongation, Yc Stress a t 300Yc elongation 25-minute cure a t 292’ F. lb.;’sq. inch 50-minute cure a t 292O F. lb./sq. inch 90-minute cure a t 292’ F. lb./sq. inch
0 5 0 8
Total Range 0 10 0 60 0.25 0 05 0.10
Canadian Synthetic Rubber, Ltd. Copolymer Corporation E. I. du Pont de Kemours & Co., Inc. Firestone Tire and Rubber CompanlGeneral Tire and Rubber Company The B. F. Goodrich Company Goodyear Synthetic Rubber Corporation Government Evaluation Laboratory Humble Oil Company National Bureau of Standards National Synthetic Rubber Corporatioil Navy Department, Bureau of Ships Office of Rubber Reserve, RFC Standard Oil Company of New Jersey United Carbon Company United States Rubber Company University of Illinois
2.5
4 0
150 20
iS0
20 25 25
100 125 125
100
peated daily tests of the same blend of GR-Susing the same lots of compounding ingredients within limits given in Table IV. The range is taken as five times the standard deviation, since 99 out of 100 values should fall within it. The average range for monthly plant data is approximately four times the standard deviation, which corresponds to a 95 out of 100 expectation. The precision of testing natural and types of synthetic rubbers other than GR-Sshould be as good as or better than that of testing GR-S. On the other hand, the accuracy of testing any rubber cannot be determined conveniently by a single laboratory because of possible systematic errors in equipment or procedure. There should, therefore, be reference lots of rubbers available to the rubber industry for detecting systematic errors in the rubber laboratories. There should also be required standard lots of compounding ingredients for preparing test vulcanizates. ACKNOWLEDGMEKT
,The accomplishments in testing reported herein were the result of the collaborative effort of groups in several laboratories. Although it is not possible to acknowledge separately each indi-
LITERATURE CITED
Holt, W. L., Knox, E. O., and Roth, F. L., J . Research S a t l . Bur. Standards (in press). Hucks, W. R., A.S.T.M. Symposium on Rubber Testing, Special Tech. Pub.74, 9 (1947). Meuser, L., Stiehler, R. D., and Hackett, R. W., I n d i a Rubber W o r l d , 117, 57 (1947). Office of Rubber Reserve, Reconstruction Finance Corp., “Speci’fications for Government Synthetic Rubbers,” pp. 23-46 (Jan. 1, 1947); amendments of April 1 and Aug. 1, 1947. Schade. J. W., and Roth. F. L.. I n d i a Rubber World, 116, 777 (1947). .
I
Taylor, R. H., Fielding, J. H., and Mooney, M.,Rubber Age (N. Y.), 61, 567, 705 (1947). Tyler, W. P., and Higuchi, T., I n d i a Rubber W o r l d , 116, 636 (1947).
RECEIVEDSeptember 18, 1947. Presented before the Division of Rubber CHEMICALSOCIETY. Chemistry a t the 112th Meeting of the AXERIC.~S New York, S. Y.
tmission spectroscopy in an Oli laboratory I
.
I
A
R . G. RUSSELL, Gulf Research & Deuelopment Company, Pittsburgh, Pa. The use of the spectrograph is different in the oil laboratory than in the metallurgical laboratory. It serves many functions not handled by any other analytical tool. In spite of the wide variety of ‘materials handled, it enables the operator quiclcly to analyze most types of materials with an accuracy in niost cases equal to that of the best chemical methods. It cannot be used indiscriminately, but usually methods can be developed for the elements i t detects.
AS? articles during the past ten years have described methods of analysis using a spectrograph-direct current arc, alternating current arc, and alternating current spark niethods, with many variations. Most have stressed the fact that the results are obtained more quickly, and n i t h an equal or better accuracy than by other methods. Such procedures may be turned over to relatively untrained personnel without loss of either accuracy or speed. While these facts are true, sight has been lost of the need in industry for a n analytical tool that is capable of giving information not otherwise obtainable, and of certain advantages that may sometimes be obtained by the use of a fairly long spectrographic procedure. It is with the above situation in mind t h a t this paper has been prepared. The methods presented are not the only solution to the problem under discussion, but are one satisfactory answer to a heterogeneous group of problems. A Bausch & Lomb large Littrow spectrograph, A.R.L. Dietert
source units, briquetting press, densitometer, and developing equipment make up the main facilities of the laboratory. QUALITATIVE A Y A L l S I S
Direct Current Arc Techniques. I t is possible to identify approximately 70 elements of the periodic table by means of an ordinary emission spectrograph. When there is interest in any of these 70 elements, most samples, regardless of physical state, are sent to t,he spectrographic laboratory for a qualitative analysis before any other work is performed on them. The working procedure is set up to give a semiquantitative as well as qualitative analysis. Samplingis particularlyimportant in view of the heterogeneityof many oY thesamplessubnlitted. Illoils, liquidsamples,or organic residues are either evaporated to dryness or ignited and burned. The residue in the case of organic materials is heated to approxi-
